The National Cancer Institute has sponsored three randomized controlled trials of screening for early lung cancer in large, high-risk populations to determine whether (1) lung cancer detection can be improved by adding sputum cytological screening every 4 months to chest roentgenography done either yearly or every 4 months; and (2) lung cancer mortality can be significantly reduced by this type of screening program, followed by appropriate treatment. Results of the three trials suggest that (1) sputum cytology alone detects 15% to 20% of lung cancers, almost all of which are squamous cancers with a favorable prognosis; and (2) chest roentgenography may be a more effective test for early-stage lung cancer than previous reports have suggested. Neverthless, results of the randomized trial conducted at the Mayo Clinic showed that offering both procedures to high-risk outpatients every 4 months conferred no mortality advantage over standard medical practice that included recommended annual testing.